
import pandas as pd
import matplotlib.pyplot as plt
import numpy as np
from scipy.stats import binom
from scipy.stats import norm
from scipy.stats import t


raw_data = pd.read_csv('lending_club_loan_two.csv')
raw_data.head(3)


raw_data.columns.tolist()

['loan_amnt',
 'term',
 'int_rate',
 'installment',
 'grade',
 'sub_grade',
 'emp_title',
 'emp_length',
 'home_ownership',
 'annual_inc',
 'verification_status',
 'issue_d',
 'loan_status',
 'purpose',
 'title',
 'dti',
 'earliest_cr_line',
 'open_acc',
 'pub_rec',
 'revol_bal',
 'revol_util',
 'total_acc',
 'initial_list_status',
 'application_type',
 'mort_acc',
 'pub_rec_bankruptcies',
 'address']


n_variables = raw_data.select_dtypes(include=["number"]).columns.tolist()
raw_data[n_variables].head(4)


drop_num_var = ['mort_acc', 'revol_bal']


c_variables = raw_data.select_dtypes(include=["object"]).columns.tolist()
raw_data[c_variables].head(4)


drop_cat_var = ['grade','emp_title', 'emp_length', 'issue_d', 
                'title', 'earliest_cr_line', 'initial_list_status', 
                'address', 'application_type']


drop_col = drop_num_var + drop_cat_var #add 2 lists of numeric and categorical variables together
drop_col

['mort_acc',
 'revol_bal',
 'grade',
 'emp_title',
 'emp_length',
 'issue_d',
 'title',
 'earliest_cr_line',
 'initial_list_status',
 'address',
 'application_type']


data_initial = raw_data.drop(labels=drop_col, axis=1) #data with columns we'd like to look at
data_initial.head(4)


data_initial.isnull().sum() / len(data_initial) * 100

loan_amnt               0.000000
term                    0.000000
int_rate                0.000000
installment             0.000000
sub_grade               0.000000
home_ownership          0.000000
annual_inc              0.000000
verification_status     0.000000
loan_status             0.000000
purpose                 0.000000
dti                     0.000000
open_acc                0.000000
pub_rec                 0.000000
revol_util              0.069692
total_acc               0.000000
pub_rec_bankruptcies    0.135091
dtype: float64


data_initial.dropna(inplace=True) #removes NaN values from data, does not create a new dataset


data_initial.isnull().sum()

loan_amnt               0
term                    0
int_rate                0
installment             0
sub_grade               0
home_ownership          0
annual_inc              0
verification_status     0
loan_status             0
purpose                 0
dti                     0
open_acc                0
pub_rec                 0
revol_util              0
total_acc               0
pub_rec_bankruptcies    0
dtype: int64


c_var = data_initial.select_dtypes(include=["object"])
c_var.head(3)


n_var = data_initial.select_dtypes(include=["number"])
n_var.head(3)


numeric_describe = pd.DataFrame({
    
    'loan_amnt' : data_initial['loan_amnt'].describe(),
    'int_rate' : data_initial['int_rate'].describe(),
    'installment' : data_initial['installment'].describe(),
    'annual_inc' : data_initial['annual_inc'].describe(),
    'dti' : data_initial['dti'].describe(),
    'pub_rec' : data_initial['pub_rec'].describe(),
    'revol_util' : data_initial['revol_util'].describe(),
    'total_acc' : data_initial['total_acc'].describe(),
    'pub_rec_bankruptcies' : data_initial['pub_rec_bankruptcies'].describe()
})


pd.options.display.float_format = '{:.2f}'.format
numeric_describe


data_initial.shape

(395219, 16)


#lets add our dti range

dti_25 = data_initial['dti'].quantile(q=0.25) #25th percentile
dti_75 = data_initial['dti'].quantile(q=0.75) #75th percentile
iqr_dti = dti_75 - dti_25

dti_lower = dti_25 - 1.5*iqr_dti #debt to income ratio lower limit
dti_upper = dti_75 + 1.5*iqr_dti #debt to income ratio upper limit

dti_range = (data_initial['dti'] <= dti_upper) & (data_initial['dti'] >= dti_lower)


data = data_initial[dti_range] #remove unwanted dti
data.shape

(394944, 16)


#lets remove unwanted total account values from the dataset

total_acc_25 = data_initial['total_acc'].quantile(q=0.25) #25th percentile
total_acc_75 = data_initial['total_acc'].quantile(q=0.75) #75th percentile
iqr_total_acc = total_acc_75 - total_acc_25

total_acc_lower = total_acc_25 - 1.5*iqr_total_acc #total accounts lower limit
total_acc_upper = total_acc_75 + 1.5*iqr_total_acc #total accounts upper limit

total_acc_range = (data_initial['total_acc'] <= total_acc_upper) & (data_initial['total_acc'] >= total_acc_lower)


data = data_initial[dti_range & total_acc_range] #remove unwanted dti and total accounts
data.shape

(386462, 16)


#lets add our annual income range

annual_inc_25 = data_initial['annual_inc'].quantile(q=0.25) #25th percentile
annual_inc_75 = data_initial['annual_inc'].quantile(q=0.75) #75th percentile
iqr_annual_inc = annual_inc_75 - annual_inc_25

annual_inc_lower = annual_inc_25 - 1.5*iqr_annual_inc #debt to income ratio lower limit
annual_inc_upper = annual_inc_75 + 1.5*iqr_annual_inc #debt to income ratio upper limit

annual_inc_range = (data_initial['annual_inc'] <= annual_inc_upper) & (data_initial['annual_inc'] >= annual_inc_lower)


data = data_initial[dti_range & annual_inc_range & total_acc_range] #remove unwanted dti, total_acc & annual_inc
data.shape

(370708, 16)


data['home_ownership'].value_counts()

MORTGAGE    181596
RENT        153636
OWN          35340
OTHER          104
NONE            29
ANY              3
Name: home_ownership, dtype: int64


data = data.loc[data['home_ownership'] != 'ANY']
data.shape

(370705, 16)


x = 370705/ 396030 * 100
print('The size of our sample is ' + '{0:.2f}%'.format(x) + ' of the entire dataframe!')

The size of our sample is 93.61% of the entire dataframe!


data.columns

Index(['loan_amnt', 'term', 'int_rate', 'installment', 'sub_grade',
       'home_ownership', 'annual_inc', 'verification_status', 'loan_status',
       'purpose', 'dti', 'open_acc', 'pub_rec', 'revol_util', 'total_acc',
       'pub_rec_bankruptcies'],
      dtype='object')


data['term'].value_counts()

 36 months    283612
 60 months     87093
Name: term, dtype: int64


labels = ['36 months', '60 months']
explode = [0, 0.1]

plt.figure(figsize=(8,8))

plt.style.use("ggplot")
plt.pie(data['term'].value_counts(),
        labels=labels,
       wedgeprops={'edgecolor' : 'black'},
        explode=explode,
       shadow=True,
       startangle=90,
       autopct='%1.1f%%', textprops = { "fontsize" : 35 })

plt.title('Loans split by months', bbox={'facecolor':'0.9', 'pad':5}, fontsize = 35) #increase font size!

plt.show()


months_36 = data.loc[data['term'] == ' 36 months']
months_60 = data.loc[data['term'] == ' 60 months']


months_36['loan_status'].value_counts()

Fully Paid     238301
Charged Off     45311
Name: loan_status, dtype: int64


months_60['loan_status'].value_counts()

Fully Paid     58838
Charged Off    28255
Name: loan_status, dtype: int64


fig, (ax1,ax2) = plt.subplots(1,2,figsize=(20,20)) #ax1,ax2 refer to two pies

labels = ['Fully Paid', 'Charged Off']
explode = [0, 0.1]

plt.style.use("ggplot")

ax1.pie(months_36['loan_status'].value_counts(),
        labels=labels,
       wedgeprops={'edgecolor' : 'black'},
        explode=explode,
       shadow=True,
       startangle=90,
       autopct='%1.1f%%', textprops = { "fontsize" : 35 })

ax1.set_title("Loan Status of 36 months Loans", bbox={'facecolor':'0.9', 'pad':5}, fontsize = 35)

ax2.pie(months_60['loan_status'].value_counts(),
        labels=labels,
       wedgeprops={'edgecolor' : 'black'},
        explode=explode,
       shadow=True,
       startangle=90,
       autopct='%1.1f%%', textprops = { "fontsize" : 35 })

ax2.set_title("Loan Status of 60 months Loans",bbox={'facecolor':'0.9', 'pad':5}, fontsize = 35 )

plt.show()


co = data[data['loan_status'] == 'Charged Off']
fp = data[data['loan_status'] == 'Fully Paid']

co_36 = months_36[months_36['loan_status'] == 'Charged Off']
co_60 = months_60[months_60['loan_status'] == 'Charged Off']

fp_36 = months_36[months_36['loan_status'] == 'Fully Paid']
fp_60 = months_60[months_60['loan_status'] == 'Fully Paid']


def bins(x_lower, x_upper, bins):
    
    """ This function manually returns bins of a variable of length 'bins' """
    
    b = ((x_upper - x_lower + 1) / bins ) #b is the interval of the bins

    bins_list = []
    for i in range(bins+1):
    
        x = x_lower

        x += ( i * b )
    
        if x >= x_upper:
            break
        else:
            bins_list.append(x)
            
        a = np.around(bins_list, 2)
        
    return a


dti_60_bins = bins(x_lower=co_60['dti'].min(), x_upper=co_60['dti'].max(), bins=25)

dti_60_bins

array([ 0.  ,  1.66,  3.32,  4.98,  6.64,  8.3 ,  9.96, 11.62, 13.28,
       14.94, 16.6 , 18.26, 19.92, 21.58, 23.24, 24.9 , 26.56, 28.22,
       29.88, 31.54, 33.2 , 34.86, 36.52, 38.18, 39.84])


def co_bins_portion(bins, data, variable, co_data):
    
    """This function returns percentage of charged off loans based on bins of a parameter of the bins"""
    
    dti_co = []
    
    for index in range(len(bins)):
    
        if index == (len(bins) - 1):
            break       
    
        else:
            range_dti = (data[variable] >= bins[index]) & (data[variable] <= bins[index + 1])
            x = len(co_data.loc[range_dti, variable]) / len(data.loc[range_dti, variable]) * 100
    
        dti_co.append(x)
    
        co_dti = [round(num, 2) for num in dti_co]

    return np.array(co_dti)


#Percentage of 60 months charged off loans based on bins of dti
co_60_dti = co_bins_portion(bins=dti_60_bins, data=months_60, 
                variable='dti', co_data=co_60)
co_60_dti

array([23.57, 20.51, 24.39, 23.57, 24.33, 25.37, 26.15, 27.86, 27.26,
       28.65, 30.32, 32.27, 33.33, 34.99, 36.4 , 37.43, 40.76, 41.22,
       43.09, 44.42, 45.44, 48.88, 52.37, 51.92])


x = np.arange(len(dti_60_bins)-1)  # the label locations
space = 0.22# the width of the bars
width = 0.9
y = np.array(co_60_dti)

plt.figure(figsize=(18,7))

plt.bar(x, y, color='r', width=width, alpha=0.6, label='Charged Off')

plt.xticks(x-0.5, dti_60_bins[:-1], fontsize=15, rotation=45)
plt.yticks(fontsize = 20)
plt.legend(fontsize=15,bbox_to_anchor=(0.2, 1.0))
plt.title('% of charged off 60 mth loans based on bins of dti', fontsize=30)

for i in range(len(y)):
    plt.text(i, y[i], y[i], ha='center', va='bottom', fontsize=13)

plt.ylabel('Frequency(%)',fontsize=30)
plt.xlabel('Debt to Income Ratio',fontsize=30)
plt.show()


x = np.arange(len(dti_60_bins)-1)  # the label locations
space = 0.22# the width of the bars
width = 0.9
y = np.array(co_60_dti)

plt.figure(figsize=(18,7))

plt.bar(x, y, color='r', width=width, alpha=0.6, label='Charged Off')

plt.xticks(x-0.5, dti_60_bins[:-1], fontsize=15, rotation=45)
plt.yticks(fontsize = 20)
plt.legend(fontsize=15,bbox_to_anchor=(0.2, 1.0))
plt.title('% of charged off 60 mth loans based on bins of dti', fontsize=30)

for i in range(len(y)):
    plt.text(i, y[i], y[i], ha='center', va='bottom', fontsize=13)

plt.ylabel('Frequency(%)',fontsize=30)
plt.xlabel('Debt to Income Ratio',fontsize=30)
plt.axhline(y=45)
plt.show()


np.sort(data['open_acc'].unique()) #number of open accounts

array([ 1.,  2.,  3.,  4.,  5.,  6.,  7.,  8.,  9., 10., 11., 12., 13.,
       14., 15., 16., 17., 18., 19., 20., 21., 22., 23., 24., 25., 26.,
       27., 28., 29., 30., 31., 32., 33., 34., 35., 36., 37., 38., 39.,
       40., 41., 42., 43., 44., 45.])


open_acc_valuecount = data['open_acc'].value_counts().sort_index()
#number of open accounts in data set


open_acc_co_valuecount = co['open_acc'].value_counts().sort_index()
#order of percentage of charged off loans in ascending order of percentage


open_acc_co = (np.around(open_acc_co_valuecount / open_acc_valuecount, 2) * 100).fillna(0)
#percentage of charged off loans based on number of open accounts


acc_bins = bins(x_lower=co['open_acc'].min(), x_upper=co['open_acc'].max(), bins=9)
acc_bins #bins of number of open accounts

array([ 1.,  6., 11., 16., 21., 26., 31., 36., 41.])


open_acc_valuecount_60 = months_60['open_acc'].value_counts().sort_index()
open_acc_co_valuecount_60 = co_60['open_acc'].value_counts().sort_index()
open_acc_co_60 = (np.around(open_acc_co_valuecount_60 / open_acc_valuecount_60, 2) * 100).fillna(0)
acc_bins_60 = np.around(bins(x_lower=co_60['open_acc'].min(), x_upper=co_60['open_acc'].max(), bins=9),0)
y_no_bankruptcy_60 = co_bins_portion(acc_bins_60, months_60, 'open_acc', co_60) 


x = np.arange(len(acc_bins_60)-1)  # the label locations
space = 0.22# the width of the bars
width = 0.9
y = np.array(y_no_bankruptcy_60)

plt.figure(figsize=(10,7))

plt.bar(x, y, color='r', width=width, alpha=0.6, label='Charged Off') 
#plots graph of percentage of charged off loans based on bins of number of open accounts

plt.xticks(x-0.4, acc_bins[:-1], fontsize=15, rotation=45)
plt.yticks(fontsize = 15)
plt.legend(fontsize=15,bbox_to_anchor=(1, 0.5))
plt.title('% of charged off 60 mths loans based on bins of Open Accounts', fontsize=18)

for i in range(len(y)):
    plt.text(i, y[i], y[i], ha='center', va='bottom', fontsize=10)
    
plt.ylim(0,45)
plt.ylabel('Frequency(%)',fontsize=20)
plt.xlabel('Number of Open Accounts',fontsize=20)
plt.show()


inc_bins = np.around(bins(x_lower=co['annual_inc'].min(), x_upper=co['annual_inc'].max(), bins=25), 0)
inc_bins

array([  4000.,  10140.,  16280.,  22420.,  28560.,  34700.,  40840.,
        46980.,  53120.,  59260.,  65400.,  71540.,  77680.,  83821.,
        89961.,  96101., 102241., 108381., 114521., 120661., 126801.,
       132941., 139081., 145221., 151361.])


inc_bins_60 = np.around(bins(x_lower=co_60['annual_inc'].min(), x_upper=co_60['annual_inc'].max(), bins=25), 0)
#annual inc bins for 60 months loans


co_bins_portion(bins=inc_bins_60, data=months_60, variable='annual_inc', co_data=co_60)

array([38.46, 37.18, 35.19, 42.13, 43.1 , 41.53, 40.07, 36.24, 35.4 ,
       34.44, 31.83, 29.84, 29.64, 28.08, 26.39, 26.95, 25.38, 24.09,
       27.56, 24.77, 24.13, 21.29, 24.03, 24.93])


#Percentage of charged off loans based on bins of annual income for 60 months loans
co_ann_inc_60 = co_bins_portion(bins=inc_bins_60, data=months_60, variable='annual_inc', co_data=co_60)


#plot graph for 60 months loans
x = np.arange(len(inc_bins_60)-1)  # the label locations
space = 0.22# the width of the bars
width = 0.9
y = np.array(co_ann_inc_60)

plt.figure(figsize=(18,7))

plt.bar(x, y, color='r', width=width, alpha=0.6, label='Charged Off')

plt.xticks(x-0.5, inc_bins[:-1], fontsize=15, rotation=45)
plt.yticks(fontsize = 20)
plt.legend(fontsize=15,bbox_to_anchor=(1.0, 1.0))
plt.title('% of charged off 60 months loans based on bins of Annual Income', fontsize=20)

for i in range(len(y)):
    plt.text(i, y[i], y[i], ha='center', va='bottom', fontsize=10)

plt.ylabel('Frequency(%)',fontsize=20)
plt.xlabel('Annual Income',fontsize=20) #convert to whole numbers
plt.show()


co = data[data['loan_status'] == 'Charged Off']
fp = data[data['loan_status'] == 'Fully Paid']


purpose_fp_order = np.array(fp['purpose'].value_counts().index) 
#order of value counts


purpose_fp = (np.around(fp['purpose'].value_counts()[purpose_fp_order]
                        / data['purpose'].value_counts()[purpose_fp_order], 2)* 100).sort_values()
#percentage of fully paid loans in order of purpose


purpose_co = np.around( 100 - purpose_fp, 2) 
#percentage of charged of loans in order of purpose


debt = (data['purpose'] == 'debt_consolidation') | (data['purpose'] == 'credit_card')

events = ((data['purpose'] == 'wedding') | (data['purpose'] == 'vacation') | (data['purpose'] == 'other'))

necessity = ((data['purpose'] == 'medical') | (data['purpose'] == 'educational'))

business_related = ((data['purpose'] == 'renewable_energy') | (data['purpose'] == 'major_purchase') 
                            | (data['purpose'] == 'small_business'))
                            
house = ((data['purpose'] == 'house') & (data['purpose'] == 'moving') 
                            | (data['purpose'] == 'home_improvement') | (data['purpose'] == 'car'))


charged_off =  data['loan_status'] == 'Charged Off'

co_debt = data.loc[debt & charged_off]
#charged off debt purpose

co_events = data.loc[events & charged_off]
#charged off events purpose

co_necessity = data.loc[necessity & charged_off]
#charged off necessity purpose

co_business_related = data.loc[business_related & charged_off]
#charged off business related purpose

co_house = data.loc[house & charged_off]
#charged off house purpose


#bins of annual income for debt purpose
debt_inc_bins = np.rint(bins(x_lower=co_debt['annual_inc'].min(),
                                 x_upper=co_debt['annual_inc'].max(), bins=25))

#bins of annual income for events purpose
events_inc_bins = np.rint(bins(x_lower=co_events['annual_inc'].min(), 
                                 x_upper=co_events['annual_inc'].max(), bins=25))

#bins of annual income for necessity purpose
necessity_inc_bins = np.rint(bins(x_lower=co_necessity['annual_inc'].min(), 
                                 x_upper=co_necessity['annual_inc'].max(), bins=25))

#bins of annual income for business related purpose
business_related_inc_bins = np.rint(bins(x_lower=co_business_related['annual_inc'].min(), 
                                 x_upper=co_business_related['annual_inc'].max(), bins=25))

#bins of annual income for house purpose
house_inc_bins = np.rint(bins(x_lower=co_house['annual_inc'].min(), 
                                 x_upper=co_house['annual_inc'].max(), bins=25))


y_debt = co_bins_portion(debt_inc_bins, data[debt], 'annual_inc', co_debt) 
#% of charged off loans for debt purpose based on bins of annual income

y_events = co_bins_portion(events_inc_bins, data[events], 'annual_inc', co_events) 
#% of charged off loans for events purpose based on bins of annual income

y_necessity = co_bins_portion(necessity_inc_bins, data[necessity], 'annual_inc', co_necessity) 
#% of charged off loans for necessity purpose based on bins of annual income

y_business_related = co_bins_portion(business_related_inc_bins, data[business_related], 'annual_inc', co_business_related) 
#% of charged off loans for business related purpose based on bins of annual income

y_house = co_bins_portion(house_inc_bins, data[house], 'annual_inc', co_house) 
#% of charged off loans for house purpose based on bins of annual income


def home_inc(inc_bins, y_axis, title):
    
    """This function creates a bar graph of charged off loans based on bins of annual income"""
    
    x = np.arange(len(inc_bins)-1)  # the label locations
    space = 0.22# the width of the bars
    width = 0.9
    y = np.array(y_axis)
    plt.figure(figsize=(15,5))
    plt.bar(x, y, color='r', width=width, alpha=0.6, label='Charged Off')
    
    plt.xticks(x-0.5, inc_bins[:-1], fontsize=20, rotation=45)
    plt.yticks(fontsize = 20)
    plt.legend(fontsize=20,bbox_to_anchor=(1, 1.05))
    plt.title(title, fontsize=20)

    for i in range(len(y)):
        plt.text(i, y[i], y[i], ha='center', va='bottom', fontsize=10)

    plt.ylabel('Frequency(%)',fontsize=30)
    plt.xlabel('Annual Income',fontsize=30)
    plt.show()


x = data['purpose' ].value_counts().index
y = np.around(data['purpose'].value_counts(normalize=True) * 100, 2)

plt.figure(figsize=(15,7))
plt.bar(x, y, color='b', alpha=0.6) #plot graph of loans proportion based on purpose
plt.xticks(rotation=78, fontsize=18)
plt.yticks(fontsize = 25)
for i in range(len(y)):
        plt.text(i-0.5, y[i], y[i], ha='left', va='bottom', fontsize=18)
        
        
plt.ylabel('Percentage (%)', fontsize =30)
plt.xlabel('Purpose', fontsize = 30)
plt.title('loans based on Purpose', fontsize=25)
plt.show()


charged_off =  data['loan_status'] == 'Charged Off'

co_60_debt = months_60.loc[debt & charged_off]
#charged off 60 mth debt purpose

co_60_events = months_60.loc[events & charged_off]
#charged off 60 mth  events purpose

co_60_necessity = months_60.loc[necessity & charged_off]
#charged off 60 mth  necessity purpose

co_60_business_related = months_60.loc[business_related & charged_off]
#charged off 60 mth  business related purpose

co_60_house = months_60.loc[house & charged_off]
#charged off 60 mth  house purpose


#bins of annual income for debt purpose for 60mth loans
debt_inc_bins_60 = np.rint(bins(x_lower=co_60_debt['annual_inc'].min(),
                                 x_upper=co_60_debt['annual_inc'].max(), bins=25))

#bins of annual income for events purpose for 60mth loans
events_inc_bins_60 = np.rint(bins(x_lower=co_60_events['annual_inc'].min(), 
                                 x_upper=co_60_events['annual_inc'].max(), bins=25))

#bins of annual income for necessity purpose for 60mth loans
necessity_inc_bins_60 = np.rint(bins(x_lower=co_60_necessity['annual_inc'].min(), 
                                 x_upper=co_60_necessity['annual_inc'].max(), bins=25))

#bins of annual income for business related purpose for 60mth loans
business_related_inc_bins_60 = np.rint(bins(x_lower=co_60_business_related['annual_inc'].min(), 
                                 x_upper=co_60_business_related['annual_inc'].max(), bins=25))

#bins of annual income for house purpose for 60mth loans
house_inc_bins_60 = np.rint(bins(x_lower=co_60_house['annual_inc'].min(), 
                                 x_upper=co_60_house['annual_inc'].max(), bins=25))


y_debt_60 = co_bins_portion(debt_inc_bins_60, months_60[debt], 'annual_inc', co_60_debt) 
#% of charged off loans for debt purpose based on bins of annual income

y_events_60 = co_bins_portion(events_inc_bins_60, months_60[events], 'annual_inc', co_60_events) 
#% of charged off loans for events purpose based on bins of annual income

y_necessity_60 = co_bins_portion(necessity_inc_bins_60, months_60[necessity], 'annual_inc', co_60_necessity) 
#% of charged off loans for necessity purpose based on bins of annual income

y_business_related_60 = co_bins_portion(business_related_inc_bins_60, months_60[business_related], 'annual_inc', co_60_business_related) 
#% of charged off loans for business related purpose based on bins of annual income

y_house_60 = co_bins_portion(house_inc_bins_60, months_60[house], 'annual_inc', co_60_house) 
#% of charged off loans for house purpose based on bins of annual income

<ipython-input-63-23f2d680cdd5>:1: UserWarning: Boolean Series key will be reindexed to match DataFrame index.
  y_debt_60 = co_bins_portion(debt_inc_bins_60, months_60[debt], 'annual_inc', co_60_debt)
<ipython-input-63-23f2d680cdd5>:4: UserWarning: Boolean Series key will be reindexed to match DataFrame index.
  y_events_60 = co_bins_portion(events_inc_bins_60, months_60[events], 'annual_inc', co_60_events)
<ipython-input-63-23f2d680cdd5>:7: UserWarning: Boolean Series key will be reindexed to match DataFrame index.
  y_necessity_60 = co_bins_portion(necessity_inc_bins_60, months_60[necessity], 'annual_inc', co_60_necessity)
<ipython-input-63-23f2d680cdd5>:10: UserWarning: Boolean Series key will be reindexed to match DataFrame index.
  y_business_related_60 = co_bins_portion(business_related_inc_bins_60, months_60[business_related], 'annual_inc', co_60_business_related)
<ipython-input-63-23f2d680cdd5>:13: UserWarning: Boolean Series key will be reindexed to match DataFrame index.
  y_house_60 = co_bins_portion(house_inc_bins_60, months_60[house], 'annual_inc', co_60_house)


home_inc(debt_inc_bins_60, y_debt_60, 'Charged off 60 mths Loans based on income bins for Debt purpose')
#plots graph of loans based on income bins for debt purpose

home_inc(events_inc_bins_60, y_events_60, 'Charged off 60 mths Loans based on income bins for Events purpose')
#plots graph of loans based on income bins for events purpose


home_inc(necessity_inc_bins_60, y_necessity_60, 'Charged off 60 mths Loans based on income for Necessity purpose')
#plots graph of loans based on income bins for necessity purpose


home_inc(business_related_inc_bins_60, y_business_related_60, 'Charged off 60 mths Loans based on income bins for business related purpose')
#plots graph of loans based on income bins for business related purposes


home_inc(house_inc_bins_60, y_house_60, 'Charged off 60 mths Loans based on income for House purpose')
#plots graph of loans based on income bins for house purpose


fp['home_ownership'].value_counts() #fully paid loans based on home ownership

MORTGAGE    150348
RENT        118729
OWN          27951
OTHER           89
NONE            22
Name: home_ownership, dtype: int64


data['home_ownership'].value_counts() #loans based on home ownership

MORTGAGE    181596
RENT        153636
OWN          35340
OTHER          104
NONE            29
Name: home_ownership, dtype: int64


fp_portion_home = np.around( fp['home_ownership'].value_counts() / data['home_ownership'].value_counts() * 100, 2).sort_values()
fp_portion_home #fully paid loans percentage based on home ownership

NONE       75.86
RENT       77.28
OWN        79.09
MORTGAGE   82.79
OTHER      85.58
Name: home_ownership, dtype: float64


co_portion_home = np.around( 100 - fp_portion_home)
co_portion_home #charged off loans percentage based on home ownership

NONE       24.00
RENT       23.00
OWN        21.00
MORTGAGE   17.00
OTHER      14.00
Name: home_ownership, dtype: float64


 months_60['home_ownership'].value_counts()

MORTGAGE    50479
RENT        28826
OWN          7781
NONE            4
OTHER           3
Name: home_ownership, dtype: int64


## 60M Loans Split Home Ownership types
y = [50479, 4, 7781, 28826, 3]
x = ['Mortgage', 'None', 'Own', 'Rent', 'Other']

colors = ['#EE3B3B', '#9A32CD', '#6495ED', '#00C957', '#68228B']

plt.pie(y,colors=colors, labels=x, autopct='%1.3f%%', startangle=90, pctdistance=0.8,textprops={'fontsize': 15})
#draw circle
centre_circle = plt.Circle((0,0),0.70,fc='white')
fig = plt.gcf()
fig.gca().add_artist(centre_circle)
# Equal aspect ratio ensures that pie is drawn as a circle
ax1.axis('equal')  
plt.tight_layout()
plt.title('60 Months Loans Split By Home Ownership', fontsize=20)

plt.show()


fp_60['home_ownership'].value_counts()

MORTGAGE    36178
RENT        17460
OWN          5196
NONE            3
OTHER           1
Name: home_ownership, dtype: int64


fp_60_portion_home = np.around( fp_60['home_ownership'].value_counts() / months_60['home_ownership'].value_counts() * 100, 2).sort_values()
fp_60_portion_home #fully paid 60 mths loans percentage based on home ownership

OTHER      33.33
RENT       60.57
OWN        66.78
MORTGAGE   71.67
NONE       75.00
Name: home_ownership, dtype: float64


co_60_portion_home = np.around( 100 - fp_60_portion_home)
co_60_portion_home #charged off 60 mths loans percentage based on home ownership

OTHER      67.00
RENT       39.00
OWN        33.00
MORTGAGE   28.00
NONE       25.00
Name: home_ownership, dtype: float64


rent = data['home_ownership'] == 'RENT'

own = (data['home_ownership'] == 'OWN')

mortgage = data['home_ownership'] == 'MORTGAGE'


charged_off =  data['loan_status'] == 'Charged Off'

co_rent = data.loc[rent & charged_off]

co_own = data.loc[own & charged_off]

co_mortgage = data.loc[mortgage & charged_off]


rent_inc_bins = np.rint(bins(x_lower=co_rent['annual_inc'].min(), #bins of annual income for rent homes
                                 x_upper=co_rent['annual_inc'].max(), bins=25))

own_inc_bins = np.rint(bins(x_lower=co_own['annual_inc'].min(), #bins of annual income for own homes
                                 x_upper=co_own['annual_inc'].max(), bins=25))

mortgage_related_inc_bins = np.rint(bins(x_lower=co_mortgage['annual_inc'].min(), #bins of annual income for mortgage homes
                                 x_upper=co_mortgage['annual_inc'].max(), bins=25))


y_rent = co_bins_portion(rent_inc_bins, data[rent], 'annual_inc', co_rent) 
#% of charged off loans for rent homes based on bins of annual income

y_own = co_bins_portion(own_inc_bins, data[own], 'annual_inc', co_own) 
#% of charged off loans for own homes based on bins of annual income

y_mortgage = co_bins_portion(mortgage_related_inc_bins, data[mortgage], 'annual_inc', co_mortgage)
#% of charged off loans for mortgage homes based on bins of annual income


def barplt(xticks, y1, y2, title, xlabel):
    
    """This function draws bar plots for fully paid and charged off loans based on their purpose"""
    
    plt.style.use("ggplot")

    x = np.arange(len(xticks))  # the label locations
    space = 0.2  # the width of the bars
    width = 0.4

    plt.figure(figsize=(18,7))

    plt.bar(x - space, y1, color='g', width=width, alpha=0.6, label='Fully Paid')
    plt.bar(x + space, y2, color='r', width=width, alpha=0.6, label='Charged Off')

    plt.xticks(x, xticks, fontsize=25, rotation=90)
    plt.yticks(fontsize=25)
    plt.legend()
    plt.title(title, fontsize=25)

    plt.ylabel('Frequency (%)',fontsize=23)
    plt.xlabel(xlabel,fontsize=23)
    plt.legend(fontsize=15, bbox_to_anchor=(1.1,1))

    for i in range(len(y1)):
        plt.text(i, y1[i], y1[i], ha='right', va='bottom', fontsize=15) #adds data labels
    
    for i in range(len(y2)):
        plt.text(i, y2[i], y2[i], ha='left', va='bottom', fontsize=15) #adds data labels
    

    plt.show()


barplt(xticks=fp_60_portion_home.index, y1=fp_60_portion_home, y2=co_60_portion_home, 
title='60 months Loan Status vs Home Ownership Type', xlabel='Home Ownership Type')
#percentage of charged off and fully paid loans based on ownership type


charged_off =  data['loan_status'] == 'Charged Off'

co_rent_60 = months_60.loc[rent & charged_off]

co_own_60 = months_60.loc[own & charged_off]

co_mortgage_60 = months_60.loc[mortgage & charged_off]


fp_60_portion_home = np.around( fp_60['home_ownership'].value_counts() / months_60['home_ownership'].value_counts() * 100, 2).sort_values()
fp_60_portion_home #fully paid loans percentage based on home ownership

OTHER      33.33
RENT       60.57
OWN        66.78
MORTGAGE   71.67
NONE       75.00
Name: home_ownership, dtype: float64


co_60_portion_home = np.around( 100 - fp_60_portion_home)
co_60_portion_home #charged off loans percentage based on home ownership

OTHER      67.00
RENT       39.00
OWN        33.00
MORTGAGE   28.00
NONE       25.00
Name: home_ownership, dtype: float64


rent_inc_bins_60 = np.rint(bins(x_lower=co_rent_60['annual_inc'].min(), #bins of annual income for rent homes
                                 x_upper=co_rent_60['annual_inc'].max(), bins=25))

own_inc_bins_60 = np.rint(bins(x_lower=co_own_60['annual_inc'].min(), #bins of annual income for own homes
                                 x_upper=co_own_60['annual_inc'].max(), bins=25))

mortgage_related_inc_bins_60 = np.rint(bins(x_lower=co_mortgage_60['annual_inc'].min(), #bins of annual income for mortgage homes
                                 x_upper=co_mortgage_60['annual_inc'].max(), bins=25))


y_rent_60 = co_bins_portion(rent_inc_bins_60, months_60[rent], 'annual_inc', co_rent_60) 
#% of charged off loans for rent homes based on bins of annual income

y_own_60 = co_bins_portion(own_inc_bins_60, months_60[own], 'annual_inc', co_own_60) 
#% of charged off loans for own homes based on bins of annual income

y_mortgage_60 = co_bins_portion(mortgage_related_inc_bins_60, months_60[mortgage], 'annual_inc', co_mortgage_60)
#% of charged off loans for mortgage homes based on bins of annual income

<ipython-input-84-b8615c266749>:1: UserWarning: Boolean Series key will be reindexed to match DataFrame index.
  y_rent_60 = co_bins_portion(rent_inc_bins_60, months_60[rent], 'annual_inc', co_rent_60)
<ipython-input-84-b8615c266749>:4: UserWarning: Boolean Series key will be reindexed to match DataFrame index.
  y_own_60 = co_bins_portion(own_inc_bins_60, months_60[own], 'annual_inc', co_own_60)
<ipython-input-84-b8615c266749>:7: UserWarning: Boolean Series key will be reindexed to match DataFrame index.
  y_mortgage_60 = co_bins_portion(mortgage_related_inc_bins_60, months_60[mortgage], 'annual_inc', co_mortgage_60)


home_inc(rent_inc_bins_60, y_rent_60, 'Charged off 60 mths  Loans based on income for Rent Home Ownership')
#percentage of charged off loans based on Rent home ownership type

home_inc(own_inc_bins_60, y_own_60, 'Charged off 60 mths Loans based on income for Own Home Ownership')
#percentage of charged off loans based on Own home ownership type

home_inc(mortgage_related_inc_bins_60, y_mortgage_60, 'Charged off 60 mths Loans based on income for Mortgage Home Ownership')
#percentage of charged off loans based on Mortgage home ownership type


def rent_recco(inc_bins, y_axis, title, limit):
    
    """This function creates a bar graph of charged off loans based on bins of annual income"""
    
    x = np.arange(len(inc_bins)-1)  # the label locations
    space = 0.22# the width of the bars
    width = 0.9
    y = np.array(y_axis)
    plt.figure(figsize=(15,5))
    plt.bar(x, y, color='r', width=width, alpha=0.6, label='Charged Off')
    
    plt.xticks(x-0.5, inc_bins[:-1], fontsize=20, rotation=45)
    plt.yticks(fontsize = 20)
    plt.legend(fontsize=20,bbox_to_anchor=(1, 1.05))
    plt.title(title, fontsize=20)

    for i in range(len(y)):
        plt.text(i, y[i], y[i], ha='center', va='bottom', fontsize=10)

    plt.ylabel('Frequency(%)',fontsize=30)
    plt.xlabel('Annual Income',fontsize=30)
    plt.axhline(y=limit)
    plt.show()
    
    
rent_recco(rent_inc_bins_60, y_rent_60, 'Charged off 60 mths  Loans based on income for Rent Home Ownership', 45)


##calculate average AI of rent and mortgage

rent = data.loc[data['home_ownership'] == 'RENT'] 
                
rent_60m = rent.loc[rent['term'] == ' 60 months']

rent_60m_mean = rent_60m['annual_inc'].mean()

mortgage = data.loc[data['home_ownership'] == 'MORTGAGE'] 
                
mortgage_60m = mortgage.loc[mortgage['term'] == ' 60 months']

mortgage_60m_mean = mortgage_60m['annual_inc'].mean()


print('rent avr AI:', rent_60m_mean, '\n', 'mortgage avr AI:', mortgage_60m_mean )

rent avr AI: 65923.09691944777 
 mortgage avr AI: 78225.5482588799


subgrade_order = np.array(['A1','A2','A3','A4','A5','B1','B2','B3','B4','B5',
                     'C1','C2','C3','C4','C5','D1','D2','D3','D4','D5',
                     'E1','E2','E3','E4','E5','F1','F2','F3','F4','F5',
                     'G1','G2','G3','G4','G5'])


co = data[data['loan_status'] == 'Charged Off']
fp = data[data['loan_status'] == 'Fully Paid']


subgrade_fp = (np.around(fp['sub_grade'].value_counts()[subgrade_order] / data['sub_grade'].value_counts()[subgrade_order], 2)* 100)
#percentage of fully paid months loans in order of 'subgrade order'


subgrade_co = np.around( 100 - subgrade_fp, 2) 
#percentage of charged off loans in order of 'subgrade order'


def subgrade(xticks, y1, y2, title, xlabel):
    
    """This function shows graphs of loan status based on subgrades x"""
    
    x = np.arange(len(xticks))  # the label locations
    space = 0.22# the width of the bars
    width = 0.4

    plt.figure(figsize=(18,4))

    plt.bar(x - space, y1, color='g', width=width, alpha=0.6, label='Fully Paid')
    plt.bar(x + space, y2, color='r', width=width, alpha=0.6, label='Charged Off')

    plt.xticks(x, xticks, fontsize=20, rotation=50)
    plt.yticks(fontsize=20)
    plt.legend(fontsize=20)
    plt.title(title, fontsize=20)

    plt.ylabel('Frequency(%)',fontsize=20)
    plt.xlabel(xlabel, fontsize=20)

    plt.show()


subgrade(xticks=subgrade_order, y1=subgrade_fp,
        y2=subgrade_co, title='Loan Status vs All Subgrades',
        xlabel='Subgrades')


subsetgrade_order_e = np.array(['E1','E2','E3','E4','E5','F1','F2','F3','F4','F5',
                     'G1','G2','G3','G4','G5'])
#order of subgrade e onwards


def subgrade_e(xticks, y1, y2, title):
    
    """This function shows graphs of loan status based on subgrades E"""
    
    x = np.arange(len(xticks))  # the label locations
    space = 0.22# the width of the bars
    width = 0.4

    plt.figure(figsize=(18,7))

    plt.bar(x - space, y1, color='g', width=width, alpha=0.6, label='Fully Paid')
    plt.bar(x + space, y2, color='r', width=width, alpha=0.6, label='Charged Off')
    for i in range(len(y2)):
        plt.text(i, y2[i], y2[i], ha='left', va='bottom', fontsize=13)

    plt.xticks(x, xticks, fontsize=25)
    plt.yticks(fontsize=30)
    plt.legend(fontsize=20,bbox_to_anchor=(1.1, 1.05))
    plt.title(title, fontsize=20)

    plt.ylabel('Frequency(%)',fontsize=25)
    plt.xlabel('Subgrade',fontsize=25)

    plt.show()


#percentage of fully paid loans based on subgrades E onwards
subsetgrade_fp = (np.around(data.loc[data['loan_status'] == 'Fully Paid', 'sub_grade'].value_counts()[subsetgrade_order_e] / data['sub_grade'].value_counts()[subsetgrade_order_e], 2)* 100)


#percentage of charged off loans based on subgrades E onwards
subsetgrade_co = np.around(100 - subsetgrade_fp, 1)


subgrade_e(xticks=subsetgrade_order_e, 
           y1=subsetgrade_fp,
           y2=subsetgrade_co, 
           title='Loan Status of Loans from grade E onwards')


def subgrade_e(xticks, y1, y2, title):
    
    """This function shows graphs of loan status based on subgrades E"""
    
    x = np.arange(len(xticks))  # the label locations
    space = 0.22# the width of the bars
    width = 0.4

    plt.figure(figsize=(18,7))

    plt.bar(x - space, y1, color='g', width=width, alpha=0.6, label='Fully Paid')
    plt.bar(x + space, y2, color='r', width=width, alpha=0.6, label='Charged Off')
    plt.plot([-1,15],[45,45], color='m', label='45%')
    for i in range(len(y2)):
        plt.text(i, y2[i], y2[i], ha='left', va='bottom', fontsize=13)

    plt.xticks(x, xticks, fontsize=25)
    plt.yticks(fontsize=30)
    plt.legend(fontsize=20,bbox_to_anchor=(1.1, 1.05))
    plt.title(title, fontsize=20)

    plt.ylabel('Frequency(%)',fontsize=25)
    plt.xlabel('Subgrade',fontsize=25)

    plt.show()


#percentage of fully paid loans based on subgrades E onwards
subsetgrade_fp = (np.around(data.loc[data['loan_status'] == 'Fully Paid', 'sub_grade'].value_counts()[subsetgrade_order_e] / data['sub_grade'].value_counts()[subsetgrade_order_e], 2)* 100)


subgrade_e(xticks=subsetgrade_order_e, 
           y1=subsetgrade_fp,
           y2=subsetgrade_co, 
           title='Loan Status of Loans from grade E onwards')


## inconclusive graph on DTI

plt.figure(figsize=(14,7))


plt.hist(fp['dti'], 25, color='g', alpha=0.6, label='Fully Paid')

plt.hist(co['dti'], 25, color= 'r', alpha=0.6, label='Charged Off')

plt.legend()
plt.title('Frequency of charged of w.r.t Debt to Income Ratio')
plt.ylabel('frequency', fontsize=13)
plt.xlabel('Debt to Income Ratio', fontsize=15)
plt.xticks(fontsize=10)

plt.show()


#Percentage of 36 months charged off loans based on bins of dti

dti_bins = bins(x_lower=co['dti'].min(), x_upper=co['dti'].max(), bins=25)
dti_36_bins = bins(x_lower=co_36['dti'].min(), x_upper=co_36['dti'].max(), bins=25)


co_36_dti = co_bins_portion(bins=dti_36_bins, data=months_36, 
                variable='dti', co_data=co_36)
co_36_dti

#Percentage of charged off loans based on bins of dti

co_dti = co_bins_portion(bins=dti_bins, data=data, 
                variable='dti', co_data=co)
co_dti

x = np.arange(len(dti_bins)-1)  # the label locations
space = 0.22# the width of the bars
width = 0.9
y = np.array(co_dti)

plt.figure(figsize=(18,7))

plt.bar(x, y, color='r', width=width, alpha=0.6, label='Charged Off')

plt.xticks(x-0.5, dti_bins[:-1], fontsize=11, rotation=45)
plt.legend(fontsize=15,bbox_to_anchor=(0.2, 1.0))
plt.title('% of charged off all loans based on bins of dti', fontsize=13)

for i in range(len(y)):
    plt.text(i, y[i], y[i], ha='center', va='bottom', fontsize=13)

plt.ylabel('Frequency(%)',fontsize=13)
plt.xlabel('Debt to Income Ratio',fontsize=16)
plt.show()

x = np.arange(len(dti_36_bins)-1)  # the label locations
space = 0.22# the width of the bars
width = 0.9
y = np.array(co_36_dti)

plt.figure(figsize=(18,7))

plt.bar(x, y, color='r', width=width, alpha=0.6, label='Charged Off')

plt.xticks(x-0.5, dti_36_bins[:-1], fontsize=11, rotation=45)
plt.legend(fontsize=15,bbox_to_anchor=(0.2, 1.0))
plt.title('% of charged off 36 mth loans based on bins of dti', fontsize=13)

for i in range(len(y)):
    plt.text(i, y[i], y[i], ha='center', va='bottom', fontsize=13)

plt.ylabel('Frequency(%)',fontsize=13)
plt.xlabel('Debt to Income Ratio',fontsize=16)
plt.show()


## 36 months
open_acc_valuecount_36 = months_36['open_acc'].value_counts().sort_index()

open_acc_co_valuecount_36 = co['open_acc'].value_counts().sort_index()
#order of percentage of charged off 36 mths loans in ascending order of percentage

open_acc_co_36 = (np.around(open_acc_co_valuecount_36 / open_acc_valuecount_36, 2) * 100).fillna(0)
                  
acc_bins_36 = np.around(bins(x_lower=co_36['open_acc'].min(), x_upper=co_36['open_acc'].max(), bins=9),0)

                  
y_no_bankruptcy_36 = co_bins_portion(acc_bins_36, months_36, 'open_acc', co_36) 
#% of charged off 36 mths loans based on bins of open accounts
                  
x = np.arange(len(acc_bins_36)-1)  # the label locations
space = 0.22# the width of the bars
width = 0.9
y = np.array(y_no_bankruptcy_36)

plt.figure(figsize=(10,7))

plt.bar(x, y, color='r', width=width, alpha=0.6, label='Charged Off') 
#plots graph of percentage of charged off loans based on bins of number of open accounts

plt.xticks(x-0.4, acc_bins[:-1], fontsize=11, rotation=45)
plt.legend(fontsize=15,bbox_to_anchor=(1, 0.5))
plt.title('% of charged off 36 mths loans based on bins of Open Accounts', fontsize=13)

for i in range(len(y)):
    plt.text(i, y[i], y[i], ha='center', va='bottom', fontsize=10)

plt.ylabel('Frequency(%)',fontsize=13)
plt.xlabel('Number of Open Accounts',fontsize=13)
plt.ylim(0,45)
plt.show()


## all loans


y_open_acc = co_bins_portion(acc_bins, data, 'open_acc', co)


x = np.arange(len(acc_bins)-1)  # the label locations
space = 0.22# the width of the bars
width = 0.9
y = np.array(y_open_acc)

plt.figure(figsize=(10,7))

plt.bar(x, y, color='r', width=width, alpha=0.6, label='Charged Off') 
#plots graph of percentage of charged off loans based on bins of number of open accounts

plt.xticks(x-0.4, acc_bins[:-1], fontsize=11, rotation=45)
plt.legend(fontsize=15,bbox_to_anchor=(1, 0.5))
plt.title('% of charged off all loans based on bins of Open Accounts', fontsize=13)

for i in range(len(y)):
    plt.text(i, y[i], y[i], ha='center', va='bottom', fontsize=10)

plt.ylabel('Frequency(%)',fontsize=13)
plt.xlabel('Number of Open Accounts',fontsize=13)
plt.show()


plt.figure(figsize=(18,7))


plt.hist(fp['annual_inc'],30, color='g', alpha=0.6, label='Fully Paid')

plt.hist(co['annual_inc'],30, color= 'r', alpha=0.6, label='Charged Off')

plt.legend()
plt.title('Frequency of charged of w.r.t Annual Income')
plt.ylabel('frequency', fontsize=13)
plt.xlabel('Annual Income', fontsize=15)
plt.xticks(fontsize=10)

plt.show() #this graph is redundant


co_bins_portion(bins=inc_bins_60, data=months_60, variable='annual_inc', co_data=co_60)

inc_bins_36 = np.around(bins(x_lower=co_36['annual_inc'].min(), x_upper=co_36['annual_inc'].max(), bins=25), 0)
#Percentage of charged off loans based on bins of annual income
co_ann_inc = co_bins_portion(bins=inc_bins, data=data, 
                variable='annual_inc', co_data=co)

#Percentage of charged off loans based on bins of annual income for 36 months loans
co_ann_inc_36 = co_bins_portion(bins=inc_bins_36, data=months_36, variable='annual_inc', co_data=co_36)



#plot graph for all loans
x = np.arange(len(inc_bins)-1)  # the label locations
space = 0.22# the width of the bars
width = 0.9
y = np.array(co_ann_inc)

plt.figure(figsize=(18,7))

plt.bar(x, y, color='r', width=width, alpha=0.6, label='Charged Off')

plt.xticks(x-0.5, inc_bins[:-1], fontsize=11, rotation=45)
plt.legend(fontsize=15,bbox_to_anchor=(1.0, 1.0))
plt.title('% of charged off all loans based on bins of Annual Income', fontsize=13)

for i in range(len(y)):
    plt.text(i, y[i], y[i], ha='center', va='bottom', fontsize=10)

plt.ylabel('Frequency(%)',fontsize=13)
plt.xlabel('Annual Income',fontsize=13) #convert to whole numbers
plt.show()

#plot graph for 36 months loans
x = np.arange(len(inc_bins_36)-1)  # the label locations
space = 0.22# the width of the bars
width = 0.9
y = np.array(co_ann_inc_36)

plt.figure(figsize=(18,7))

plt.bar(x, y, color='r', width=width, alpha=0.6, label='Charged Off')

plt.xticks(x-0.5, inc_bins_36[:-1], fontsize=11, rotation=45)
plt.legend(fontsize=15,bbox_to_anchor=(1.0, 1.0))
plt.title('% of charged off 36 months loans based on bins of Annual Income', fontsize=13)

for i in range(len(y)):
    plt.text(i, y[i], y[i], ha='center', va='bottom', fontsize=10)

plt.ylabel('Frequency(%)',fontsize=13)
plt.xlabel('Annual Income',fontsize=13) #convert to whole numbers
plt.show()


def barplt(xticks, y1, y2, title, xlabel):
    
    """This function draws bar plots for fully paid and charged off loans based on their purpose"""
    
    plt.style.use("ggplot")

    x = np.arange(len(xticks))  # the label locations
    space = 0.2  # the width of the bars
    width = 0.4

    plt.figure(figsize=(18,7))

    plt.bar(x - space, y1, color='g', width=width, alpha=0.6, label='Fully Paid')
    plt.bar(x + space, y2, color='r', width=width, alpha=0.6, label='Charged Off')

    plt.xticks(x, xticks, fontsize=25, rotation=90)
    plt.yticks(fontsize=25)
    plt.legend()
    plt.title(title, fontsize=25)

    plt.ylabel('Frequency (%)',fontsize=23)
    plt.xlabel(xlabel,fontsize=23)
    plt.legend(fontsize=15, bbox_to_anchor=(1.1,1))

    for i in range(len(y1)):
        plt.text(i, y1[i], y1[i], ha='right', va='bottom', fontsize=15) #adds data labels
    
    for i in range(len(y2)):
        plt.text(i, y2[i], y2[i], ha='left', va='bottom', fontsize=15) #adds data labels
    

    plt.show()
    
    
barplt(xticks=purpose_co.index, #plot graph of charged off & fully paid proportion based on purpose
       y1=purpose_fp, 
       y2=purpose_co, 
       title='Proportion of Charged off loans based on purpose (in ascending order)', xlabel='Purpose')


home_inc(necessity_inc_bins, y_necessity, 'Charged off Loans based on income for Necessity purpose')
#plots graph of loans based on income bins for necessity purpose

home_inc(debt_inc_bins, y_debt, 'Charged off Loans based on income bins for Debt purpose')
#plots graph of loans based on income bins for debt purpose

home_inc(events_inc_bins, y_events, 'Charged off Loans based on income bins for Events purpose')
#plots graph of loans based on income bins for events purpose

home_inc(business_related_inc_bins, y_business_related, 'Charged off Loans based on income bins for business related purpose')
#plots graph of loans based on income bins for business related purposes

home_inc(house_inc_bins, y_house, 'Charged off Loans based on income for House purpose')
#plots graph of loans based on income bins for house purpose


charged_off =  data['loan_status'] == 'Charged Off'

co_36_debt = months_36.loc[debt & charged_off]
#charged off 36 mth debt purpose

co_36_events = months_36.loc[events & charged_off]
#charged off 36 mth  events purpose

co_36_necessity = months_36.loc[necessity & charged_off]
#charged off 36 mth  necessity purpose

co_36_business_related = months_36.loc[business_related & charged_off]
#charged off 36 mth  business related purpose

co_36_house = months_36.loc[house & charged_off]
#charged off 36 mth  house purpose


#bins of annual income for debt purpose for 36mth loans
debt_inc_bins_36 = np.rint(bins(x_lower=co_36_debt['annual_inc'].min(),
                                 x_upper=co_36_debt['annual_inc'].max(), bins=25))

#bins of annual income for events purpose for 36mth loans
events_inc_bins_36 = np.rint(bins(x_lower=co_36_events['annual_inc'].min(), 
                                 x_upper=co_36_events['annual_inc'].max(), bins=25))

#bins of annual income for necessity purpose for 36mth loans
necessity_inc_bins_36 = np.rint(bins(x_lower=co_36_necessity['annual_inc'].min(), 
                                 x_upper=co_36_necessity['annual_inc'].max(), bins=25))

#bins of annual income for business related purpose for 36mth loans
business_related_inc_bins_36 = np.rint(bins(x_lower=co_36_business_related['annual_inc'].min(), 
                                 x_upper=co_36_business_related['annual_inc'].max(), bins=25))

#bins of annual income for house purpose for 36mth loans
house_inc_bins_36 = np.rint(bins(x_lower=co_36_house['annual_inc'].min(), 
                                 x_upper=co_36_house['annual_inc'].max(), bins=25))

y_debt_36 = co_bins_portion(debt_inc_bins_36, months_36[debt], 'annual_inc', co_36_debt) 
#% of charged off loans for debt purpose based on bins of annual income

y_events_36 = co_bins_portion(events_inc_bins_36, months_36[events], 'annual_inc', co_36_events) 
#% of charged off loans for events purpose based on bins of annual income

y_necessity_36 = co_bins_portion(necessity_inc_bins_36, months_36[necessity], 'annual_inc', co_36_necessity) 
#% of charged off loans for necessity purpose based on bins of annual income

y_business_related_36 = co_bins_portion(business_related_inc_bins_36, months_36[business_related], 'annual_inc', co_36_business_related) 
#% of charged off loans for business related purpose based on bins of annual income

y_house_36 = co_bins_portion(house_inc_bins_36, months_36[house], 'annual_inc', co_36_house) 
#% of charged off loans for house purpose based on bins of annual income

home_inc(debt_inc_bins_36, y_debt_36, 'Charged off 36 mths Loans based on income bins for Debt purpose')
#plots graph of loans based on income bins for debt purpose

home_inc(events_inc_bins_36, y_events_36, 'Charged off 36 mthsLoans based on income bins for Events purpose')
#plots graph of loans based on income bins for events purpose

home_inc(necessity_inc_bins_36, y_necessity_36, 'Charged off 36 mths Loans based on income for Necessity purpose')
#plots graph of loans based on income bins for necessity purpose

home_inc(business_related_inc_bins_36, y_business_related_36, 'Charged off 36 mths Loans based on income bins for business related purpose')
#plots graph of loans based on income bins for business related purposes

home_inc(house_inc_bins_36, y_house_36, 'Charged off 36 mths Loans based on income for House purpose')
#plots graph of loans based on income bins for house purpose

<ipython-input-109-2b22341b92cd>:39: UserWarning: Boolean Series key will be reindexed to match DataFrame index.
  y_debt_36 = co_bins_portion(debt_inc_bins_36, months_36[debt], 'annual_inc', co_36_debt)
<ipython-input-109-2b22341b92cd>:42: UserWarning: Boolean Series key will be reindexed to match DataFrame index.
  y_events_36 = co_bins_portion(events_inc_bins_36, months_36[events], 'annual_inc', co_36_events)
<ipython-input-109-2b22341b92cd>:45: UserWarning: Boolean Series key will be reindexed to match DataFrame index.
  y_necessity_36 = co_bins_portion(necessity_inc_bins_36, months_36[necessity], 'annual_inc', co_36_necessity)
<ipython-input-109-2b22341b92cd>:48: UserWarning: Boolean Series key will be reindexed to match DataFrame index.
  y_business_related_36 = co_bins_portion(business_related_inc_bins_36, months_36[business_related], 'annual_inc', co_36_business_related)
<ipython-input-109-2b22341b92cd>:51: UserWarning: Boolean Series key will be reindexed to match DataFrame index.
  y_house_36 = co_bins_portion(house_inc_bins_36, months_36[house], 'annual_inc', co_36_house)


barplt(xticks=fp_portion_home.index, y1=fp_portion_home, y2=co_portion_home, 
title='Loan Status vs Home Ownership Type', xlabel='Home Ownership Type')
#percentage of charged off and fully paid loans based on ownership type


home_inc(own_inc_bins, y_own, 'Charged off Loans based on income for Own Home Ownership')
#percentage of charged off loans based on Own home ownership type

home_inc(mortgage_related_inc_bins, y_mortgage, 'Charged off Loans based on income for Mortgage Home Ownership')
#percentage of charged off loans based on Mortgage home ownership type

home_inc(rent_inc_bins, y_rent, 'Charged off Loans based on income for Rent Home Ownership')
#percentage of charged off loans based on Rent home ownership type


data['pub_rec_bankruptcies'].value_counts()

0.00    327659
1.00     40850
2.00      1738
3.00       335
4.00        79
5.00        31
6.00         7
7.00         4
8.00         2
Name: pub_rec_bankruptcies, dtype: int64


bankruptcies_co = (np.around(co['pub_rec_bankruptcies'].value_counts() / data['pub_rec_bankruptcies'].value_counts(), 2)* 100)
#percentage of charged off loans by bankruptcies


bankruptcies_xticks = data['pub_rec_bankruptcies'].unique()


def bankruptcies(xticks, y, title):
    
    """This function shows graphs of loan status based on number of bankruptcies"""
    
    x = np.arange(len(xticks))  # the label locations
    space = 0.22# the width of the bars
    width = 0.4

    plt.figure(figsize=(18,7))

    plt.bar(x, y, color='g', width=width, alpha=0.6, label='Charged Off')
    plt.plot(x , y , color='r', alpha=0.6, label='Charged Off', marker='s')
    for i in range(len(y)):
        plt.text(i, y[i], y[i], ha='left', va='bottom', fontsize=13)

    plt.xticks(x, xticks, fontsize=11)
    plt.legend(fontsize=15,bbox_to_anchor=(1, 1.05))
    plt.title(title, fontsize=13)

    plt.ylabel('Frequency(%)',fontsize=13)
    plt.xlabel('Number of Bankruptcies',fontsize=13)

    plt.show()


no_bankruptcy = data['pub_rec_bankruptcies'] == 0

one_bankruptcy = data['pub_rec_bankruptcies'] == 1

more_than_one_bankruptcy = data['pub_rec_bankruptcies'] >= 2


charged_off =  data['loan_status'] == 'Charged Off'

co_no_bankruptcy = data.loc[no_bankruptcy & charged_off]
#charged off no bankruptcy

co_one_bankruptcy = data.loc[one_bankruptcy & charged_off]
#charged off 1 bankruptcy

co_more_than_one_bankruptcy = data.loc[more_than_one_bankruptcy & charged_off]
#charged off more than 1 bankruptcies


#bins of annual income for no bankruptcy
no_bankrupt_bins = np.rint(bins(x_lower=co_no_bankruptcy['annual_inc'].min(), 
                                x_upper=co_no_bankruptcy['annual_inc'].max(), bins=25))

#bins of annual income for one bankruptcy
one_bankruptcy_bins = np.rint(bins(x_lower=co_one_bankruptcy['annual_inc'].min(), 
                                 x_upper=co_one_bankruptcy['annual_inc'].max(), bins=25))

#bins of annual income for more than one bankruptcy
more_than_one_bankruptcy_bins = np.rint(bins(x_lower=co_more_than_one_bankruptcy['annual_inc'].min(), 
                                 x_upper=co_more_than_one_bankruptcy['annual_inc'].max(), bins=25))


y_no_bankruptcy = co_bins_portion(no_bankrupt_bins, data[no_bankruptcy], 'annual_inc', co_no_bankruptcy) 
#% of charged off loans for no bankruptcy based on bins of annual income

y_one_bankruptcy = co_bins_portion(one_bankruptcy_bins, data[one_bankruptcy], 'annual_inc', co_one_bankruptcy) 
#% of charged off loans for one bankruptcy based on bins of annual income

y_more_than_one_bankruptcy = co_bins_portion(more_than_one_bankruptcy_bins, data[more_than_one_bankruptcy], 'annual_inc', co_more_than_one_bankruptcy) 
#% of charged off loans for more than one bankruptcy based on bins of annual income


#percentage of charged off loans based on no bankruptcy
home_inc(no_bankrupt_bins, y_no_bankruptcy, 
         'Charged off Loans based on income for No Bankruptcy')

#percentage of charged off loans based on one bankruptcy
home_inc(one_bankruptcy_bins, y_one_bankruptcy, 
         'Charged off Loans based on income for One Bankruptcy')

#percentage of charged off loans based on more than one bankruptcy
home_inc(more_than_one_bankruptcy_bins, y_more_than_one_bankruptcy, 
         'Charged off Loans based on income for More than one Bankruptcy')


charged_off =  data['loan_status'] == 'Charged Off'

co_no_bankruptcy_36 = months_36.loc[no_bankruptcy & charged_off]
#charged off no bankruptcy

co_one_bankruptcy_36 = months_36.loc[one_bankruptcy & charged_off]
#charged off 1 bankruptcy

co_more_than_one_bankruptcy_36 = months_36.loc[more_than_one_bankruptcy & charged_off]
#charged off more than 1 bankruptcies


#bins of annual income for no bankruptcy
no_bankrupt_bins_36 = np.rint(bins(x_lower=co_no_bankruptcy_36['annual_inc'].min(), 
                                x_upper=co_no_bankruptcy_36['annual_inc'].max(), bins=25))

#bins of annual income for one bankruptcy
one_bankruptcy_bins_36 = np.rint(bins(x_lower=co_one_bankruptcy_36['annual_inc'].min(), 
                                 x_upper=co_one_bankruptcy_36['annual_inc'].max(), bins=25))

#bins of annual income for more than one bankruptcy
more_than_one_bankruptcy_bins_36 = np.rint(bins(x_lower=co_more_than_one_bankruptcy_36['annual_inc'].min(), 
                                 x_upper=co_more_than_one_bankruptcy_36['annual_inc'].max(), bins=25))


y_no_bankruptcy_36 = co_bins_portion(no_bankrupt_bins_36, months_36[no_bankruptcy], 'annual_inc', co_no_bankruptcy_36) 
#% of charged off loans for no bankruptcy based on bins of annual income

y_one_bankruptcy_36 = co_bins_portion(one_bankruptcy_bins_36, months_36[one_bankruptcy], 'annual_inc', co_one_bankruptcy_36) 
#% of charged off loans for one bankruptcy based on bins of annual income

y_more_than_one_bankruptcy_36 = co_bins_portion(more_than_one_bankruptcy_bins_36, months_36[more_than_one_bankruptcy], 'annual_inc', co_more_than_one_bankruptcy_36) 
#% of charged off loans for more than one bankruptcy based on bins of annual income

<ipython-input-123-c49db7bca9c5>:1: UserWarning: Boolean Series key will be reindexed to match DataFrame index.
  y_no_bankruptcy_36 = co_bins_portion(no_bankrupt_bins_36, months_36[no_bankruptcy], 'annual_inc', co_no_bankruptcy_36)
<ipython-input-123-c49db7bca9c5>:4: UserWarning: Boolean Series key will be reindexed to match DataFrame index.
  y_one_bankruptcy_36 = co_bins_portion(one_bankruptcy_bins_36, months_36[one_bankruptcy], 'annual_inc', co_one_bankruptcy_36)
<ipython-input-123-c49db7bca9c5>:7: UserWarning: Boolean Series key will be reindexed to match DataFrame index.
  y_more_than_one_bankruptcy_36 = co_bins_portion(more_than_one_bankruptcy_bins_36, months_36[more_than_one_bankruptcy], 'annual_inc', co_more_than_one_bankruptcy_36)


charged_off_60 =  data['loan_status'] == 'Charged Off'

co_no_bankruptcy_60 = months_60.loc[no_bankruptcy & charged_off]
#charged off no bankruptcy

co_one_bankruptcy_60 = months_60.loc[one_bankruptcy & charged_off]
#charged off 1 bankruptcy

co_more_than_one_bankruptcy_60 = months_60.loc[more_than_one_bankruptcy & charged_off]
#charged off more than 1 bankruptcies


#bins of annual income for no bankruptcy
no_bankrupt_bins_60 = np.rint(bins(x_lower=co_no_bankruptcy_60['annual_inc'].min(), 
                                x_upper=co_no_bankruptcy_60['annual_inc'].max(), bins=25))

#bins of annual income for one bankruptcy
one_bankruptcy_bins_60 = np.rint(bins(x_lower=co_one_bankruptcy_60['annual_inc'].min(), 
                                 x_upper=co_one_bankruptcy_60['annual_inc'].max(), bins=25))

#bins of annual income for more than one bankruptcy
more_than_one_bankruptcy_bins_60 = np.rint(bins(x_lower=co_more_than_one_bankruptcy_60['annual_inc'].min(), 
                                 x_upper=co_more_than_one_bankruptcy_60['annual_inc'].max(), bins=25))


y_no_bankruptcy_60 = co_bins_portion(no_bankrupt_bins_60, months_60[no_bankruptcy], 'annual_inc', co_no_bankruptcy_60) 
#% of charged off loans for no bankruptcy based on bins of annual income

y_one_bankruptcy_60 = co_bins_portion(one_bankruptcy_bins_60, months_60[one_bankruptcy], 'annual_inc', co_one_bankruptcy_60) 
#% of charged off loans for one bankruptcy based on bins of annual income

y_more_than_one_bankruptcy_60 = co_bins_portion(more_than_one_bankruptcy_bins_60, months_60[more_than_one_bankruptcy], 'annual_inc', co_more_than_one_bankruptcy_60) 
#% of charged off loans for more than one bankruptcy based on bins of annual income

<ipython-input-126-a1517adfc876>:1: UserWarning: Boolean Series key will be reindexed to match DataFrame index.
  y_no_bankruptcy_60 = co_bins_portion(no_bankrupt_bins_60, months_60[no_bankruptcy], 'annual_inc', co_no_bankruptcy_60)
<ipython-input-126-a1517adfc876>:4: UserWarning: Boolean Series key will be reindexed to match DataFrame index.
  y_one_bankruptcy_60 = co_bins_portion(one_bankruptcy_bins_60, months_60[one_bankruptcy], 'annual_inc', co_one_bankruptcy_60)
<ipython-input-126-a1517adfc876>:7: UserWarning: Boolean Series key will be reindexed to match DataFrame index.
  y_more_than_one_bankruptcy_60 = co_bins_portion(more_than_one_bankruptcy_bins_60, months_60[more_than_one_bankruptcy], 'annual_inc', co_more_than_one_bankruptcy_60)


def bankruptcy_inc(inc_bins, y_axis, title, ylim):
    
    """This function creates a bar graph of charged off loans based on bins of annual income"""
    
    x = np.arange(len(inc_bins)-1)  # the label locations
    space = 0.22# the width of the bars
    width = 0.9
    y = np.array(y_axis)
    plt.figure(figsize=(15,5))
    plt.bar(x, y, color='r', width=width, alpha=0.6, label='Charged Off')
    
    plt.xticks(x-0.5, inc_bins[:-1], fontsize=11, rotation=45)
    plt.legend(fontsize=15,bbox_to_anchor=(1, 1.05))
    plt.title(title, fontsize=13)
    plt.ylim(0,ylim)
    
    for i in range(len(y)):
        plt.text(i, y[i], y[i], ha='center', va='bottom', fontsize=10)

    plt.ylabel('Frequency(%)',fontsize=13)
    plt.xlabel('Annual Income')
    plt.show()


#percentage of charged off 36 mths loans based on no bankruptcy
bankruptcy_inc(no_bankrupt_bins_36, y_no_bankruptcy_36, 
         'Charged off 36 mths Loans based on income for No Bankruptcy', 45)

#percentage of charged off 60 mths loans based on no bankruptcy
bankruptcy_inc(no_bankrupt_bins_60, y_no_bankruptcy_60, 
         'Charged off 60 mths Loans based on income for No Bankruptcy', 45)

#percentage of charged off loans based on one bankruptcy
bankruptcy_inc(one_bankruptcy_bins_36, y_one_bankruptcy_36, 
         'Charged off 36 mths Loans based on income for One Bankruptcy', 100)

#percentage of charged off loans based on one bankruptcy
bankruptcy_inc(one_bankruptcy_bins_60, y_one_bankruptcy_60, 
         'Charged off 60 mths Loans based on income for One Bankruptcy', 100)

#percentage of charged off loans based on more than one bankruptcy
bankruptcy_inc(more_than_one_bankruptcy_bins_36, y_more_than_one_bankruptcy_36, 
         'Charged off 36 mths Loans based on income for More than one Bankruptcy', 80)


#percentage of charged off loans based on more than one bankruptcy
bankruptcy_inc(more_than_one_bankruptcy_bins_60, y_more_than_one_bankruptcy_60, 
         'Charged off 60 mths Loans based on income for More than one Bankruptcy', 80)

	loan_amnt	int_rate	installment	annual_inc	dti	open_acc	revol_bal	revol_util	total_acc	mort_acc
0	10000.0	11.44	329.48	117000.0	26.24	16.0	36369.0	41.8	25.0	0.0
1	8000.0	11.99	265.68	65000.0	22.05	17.0	20131.0	53.3	27.0	3.0
2	15600.0	10.49	506.97	43057.0	12.79	13.0	11987.0	92.2	26.0	0.0
3	7200.0	6.49	220.65	54000.0	2.60	6.0	5472.0	21.5	13.0	0.0

	loan_amnt	int_rate	installment	annual_inc	dti	open_acc	revol_util	total_acc
0	10000.0	11.44	329.48	117000.0	26.24	16.0	41.8	25.0
1	8000.0	11.99	265.68	65000.0	22.05	17.0	53.3	27.0
2	15600.0	10.49	506.97	43057.0	12.79	13.0	92.2	26.0

	loan_amnt	int_rate	installment	annual_inc	dti	pub_rec	revol_util	total_acc	pub_rec_bankruptcies
count	395219.00	395219.00	395219.00	395219.00	395219.00	395219.00	395219.00	395219.00	395219.00
mean	14122.06	13.64	432.07	74199.37	17.39	0.18	53.81	25.43	0.12
std	8357.05	4.47	250.70	61557.25	18.03	0.53	24.44	11.89	0.36
min	500.00	5.32	16.08	0.00	0.00	0.00	0.00	2.00	0.00
25%	8000.00	10.49	250.33	45000.00	11.30	0.00	35.90	17.00	0.00
50%	12000.00	13.33	375.49	64000.00	16.92	0.00	54.80	24.00	0.00
75%	20000.00	16.55	567.79	90000.00	22.99	0.00	72.90	32.00	0.00
max	40000.00	30.99	1533.81	8706582.00	9999.00	86.00	892.30	151.00	8.00

1.2 Data Cleaning¶

Now we're done pre-processing our data!¶

2 Data Analysis¶

Results show that longer term loans have double the proportion of loans being charged off. Hence, our team has chosen to focus our analysis on the 60 month loans¶

Numeric Variables¶

2.1 Debt to Income Ratio¶

DTI Recommendation¶

2.2 Open Accounts¶

2.3 Annual Income¶

2.4 purpose¶

group up purposes:¶

2.5 Home Ownership¶

Rent Recommendation¶

2.6 subgrades¶

Subgrade Recommendation¶

Appendix: Extra graphs and workings¶

DTI¶

DTI x Frequency of Charged Off for all loans + 36M¶

Open Accounts x Frequency of Charged Off for all loans + 36M¶

Annual Income¶

Annual Income X Frequency of Charged Off for all loans + 36M¶

Purpose¶

annual income x frequency of charged off based on all purposes (all loans)¶

annual income x frequency of charged off based on all purposes (36M)¶

Home ownership (all loans)¶

Bankruptcy (not used in report)¶

	loan_amnt	term	int_rate	installment	grade	sub_grade	emp_title	emp_length	home_ownership	annual_inc	...	open_acc	revol_bal	revol_util	total_acc	initial_list_status	application_type	mort_acc	address
0	10000.0	36 months	11.44	329.48	B	B4	Marketing	10+ years	RENT	117000.0	...	16.0	36369.0	41.8	25.0	w	INDIVIDUAL	0.0	0174 Michelle Gateway\r\nMendozaberg, OK 22690
1	8000.0	36 months	11.99	265.68	B	B5	Credit analyst	4 years	MORTGAGE	65000.0	...	17.0	20131.0	53.3	27.0	f	INDIVIDUAL	3.0	1076 Carney Fort Apt. 347\r\nLoganmouth, SD 05113
2	15600.0	36 months	10.49	506.97	B	B3	Statistician	< 1 year	RENT	43057.0	...	13.0	11987.0	92.2	26.0	f	INDIVIDUAL	0.0	87025 Mark Dale Apt. 269\r\nNew Sabrina, WV 05113

	term	sub_grade	home_ownership	verification_status	loan_status	purpose
0	36 months	B4	RENT	Not Verified	Fully Paid	vacation
1	36 months	B5	MORTGAGE	Not Verified	Fully Paid	debt_consolidation
2	36 months	B3	RENT	Source Verified	Fully Paid	credit_card